"x.py check" shows error twice

[RalfJung]

I just got this output from ./x.py check:

error[E0277]: the trait bound `&rustc_hir::def_id::DefId: rustc_middle::ty::query::sealed::IntoQueryParam<rustc_hir::def_id::DefId>` is not satisfied
   --> compiler/rustc_mir/src/transform/promote_consts.rs:748:53
    |
748 |                 if !self.tcx.is_promotable_const_fn(def_id) {
    |                                                     ^^^^^^ the trait `rustc_middle::ty::query::sealed::IntoQueryParam<rustc_hir::def_id::DefId>` is not implemented for `&rustc_hir::def_id::DefId`

error: aborting due to previous error

For more information about this error, try `rustc --explain E0277`.
error: could not compile `rustc_mir`.

To learn more, run the command again with --verbose.
warning: build failed, waiting for other jobs to finish...
error[E0277]: the trait bound `&rustc_hir::def_id::DefId: rustc_middle::ty::query::sealed::IntoQueryParam<rustc_hir::def_id::DefId>` is not satisfied
   --> compiler/rustc_mir/src/transform/promote_consts.rs:748:53
    |
748 |                 if !self.tcx.is_promotable_const_fn(def_id) {
    |                                                     ^^^^^^ the trait `rustc_middle::ty::query::sealed::IntoQueryParam<rustc_hir::def_id::DefId>` is not implemented for `&rustc_hir::def_id::DefId`

error: aborting due to previous error

For more information about this error, try `rustc --explain E0277`.
error: build failed

I also noticed that the cargo "status bar" will show things like rustc_mir, rustc_mir or rustc_driver, rustc_driver. Is it building everything twice?

Cc @Mark-Simulacrum

[ehuss]

This is likely a consequence of #76258. Cargo checks the library and unit test in parallel, which can result in duplicate messages. This is a cargo issue (tracked in rust-lang/cargo#5128). I'm thinking about adding a hash to track duplicates, but it will likely be a little tricky.

[RalfJung]

Ah that makes sense... closing in favor of rust-lang/cargo#5128 then.

[Mark-Simulacrum]

FWIW we could "fix" this by running first the regular check and then the all-target check, but that would likely be a slowdown on most systems as you lose a bunch of potential parallelism.